Chapter 7. Unsupervised Machine Learning
In the previous chapter, we learned about supervised machine learning algorithms and how we can use them in real-world scenarios.
Unsupervised learning is a little bit different and harder. The aim is to have the system learn something, but we ourselves don't know what to learn. There are two approaches to the unsupervised learning.
One approach is to find the similarities/patterns in the datasets. Then we can create clusters of these similar points. We make the assumption that the clusters that we found can be classified and can be provided with a label.
The algorithm itself cannot assign names because it doesn't have any. It can only find the clusters based on the similarities, but nothing more than that. To actually be able to find meaningful clusters, a good size of dataset is required.
It is used extensively in finding similar users, recommender systems, text classification, and so on.
We will discuss various clustering algorithms in detail. In this...
